search

zhongkaifu / RNNSharp

RNNSharp is a toolkit of deep recurrent neural network which is widely used for many different kinds of tasks, such as sequence labeling, sequence-to-sequence and so on. It's written by C# language and based on .NET framework 4.6 or above versions. RNNSharp supports many different types of networks, such as forward and bi-directional network, sequence-to-sequence network, and different types of layers, such as LSTM, Softmax, sampled Softmax and others.
BSD 3-Clause "New" or "Revised" License
285 stars 92 forks source link

OutofMemoryException #23

Closed Bilgeyar closed 7 years ago

Bilgeyar commented 8 years ago

Thanks for this greet job! I have got a problem when i start to use the new version( I used the old version on codeplex before) it throws outofmemory exception when I used old version text2vec tool generated vector.bin file.if there are a way to convert it to new version.

WordEMWrapFeaturizer funcion model.LoadBinaryModel(filename); i found that there is no bw.Write(0);//no VQ in the old version,maybe this rises the exception,is there any solutions for this,I donot want to retrain the word_vector.bin.

zhongkaifu commented 8 years ago

Yes, model format was changed for VQ which can reduce model size significantly, so Txt2Vec model is not compatible with the previous one. To avoid retraining word_vector.bin, you could revert Txt2Vec model load code (model.LoadBinaryModl(...)), and still use the previous version without VQ. And then compile RNNSharp by using modified Txt2Vec.dll, then it should work.